Ryzen 7000 Testing

[fabiosun]

Are you referring to boot argument pci=0x8000000? We discovered it here:

yep I misunderstood from a link given on AMD discord (I meant about the genesis of that boot arg)

IOPCIFamily/IOKit/pci/IOPCIConfigurator.h at IOPCIFamily-583.40.1 · apple-oss-distributions/IOPCIFamily

Contribute to apple-oss-distributions/IOPCIFamily development by creating an account on GitHub.

github.com

 

[CaseySJ]

yep I misunderstood from a link given on AMD discord (I meant about the genesis of that boot arg)

IOPCIFamily/IOKit/pci/IOPCIConfigurator.h at IOPCIFamily-583.40.1 · apple-oss-distributions/IOPCIFamily

Contribute to apple-oss-distributions/IOPCIFamily development by creating an account on GitHub.

github.com

We described all of those boot args here:

Ryzen 7000 Testing

@ExtremeXT Unfortunately pci-msi-limit did not make a difference with either 0 or 1000.

forum.amd-osx.com

kIOPCIConfiguratorDefaultETF is the second one from the end of the list.

 

[ExtremeXT]

Yes please do.

Shaneee merged it, can you do one final test to make sure it works fine?

 

[CaseySJ]

Shaneee merged it, can you do one final test to make sure it works fine?

It is 2:50am — just sipping my morning coffee in front of the fireplace. My Zen moment before the world around me wakes up.

Will retest soon.

 

[CaseySJ]

Shaneee merged it, can you do one final test to make sure it works fine?

There is a problem: base is empty:

 

[fabiosun]

@CaseySJ @ExtremeXT

some users using clover boot loader and MSI board could benefit of your work


if you have no objection I would like to put the patch in the Clover AMD Vanilla patches repo
Patch also works fine in Clover boot loader

17:514  0:001  OC: Prelinked patch [1] com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)
17:516  0:001  Replace E6117206 by E6117306 at ofs:7289
17:532  0:015  OCAK: 64-bit com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0) replace count - 1
17:533  0:001  OC: Prelinked patcher result 1 for com.apple.iokit.IOPCIFamily (com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)) - Success

thanks

 

[ExtremeXT]

There is a problem: base is empty:

View attachment 8290

It shouldn't be needed for the 4 byte patch.

 

[ExtremeXT]

@CaseySJ @ExtremeXT

some users using clover boot loader and MSI board could benefit of your work
View attachment 8291

if you have no objection I would like to put the patch in the Clover AMD Vanilla patches repo
Patch also works fine in Clover boot loader

17:514  0:001  OC: Prelinked patch [1] com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)
17:516  0:001  Replace E6117206 by E6117306 at ofs:7289
17:532  0:015  OCAK: 64-bit com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0) replace count - 1
17:533  0:001  OC: Prelinked patcher result 1 for com.apple.iokit.IOPCIFamily (com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)) - Success

thanks

I have no problem, if Casey agrees you can PR it to the repo, don't forget to edit the README as well.

 

[CaseySJ]

It shouldn't be needed for the 4 byte patch.

There’s a higher chance of conflict with future macOS updates as the number of “Find” bytes decreases. Adding ‘base’ reduces those chances. It also allows OpenCore to jump quickly to the base function (via symbol map in binary file), hence faster.

 

[CaseySJ]

@CaseySJ @ExtremeXT

some users using clover boot loader and MSI board could benefit of your work
View attachment 8291

if you have no objection I would like to put the patch in the Clover AMD Vanilla patches repo
Patch also works fine in Clover boot loader

17:514  0:001  OC: Prelinked patch [1] com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)
17:516  0:001  Replace E6117206 by E6117306 at ofs:7289
17:532  0:015  OCAK: 64-bit com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0) replace count - 1
17:533  0:001  OC: Prelinked patcher result 1 for com.apple.iokit.IOPCIFamily (com.apple.iokit.IOPCIFamily (CaseySJ - probeBusGated Disable 10 bit tags - 12.0/13.0)) - Success

thanks

No objection from me either.

 

[ExtremeXT]

There’s a higher chance of conflict with future macOS updates as the number of “Find” bytes decreases. Adding ‘base’ reduces those chances. It also allows OpenCore to jump quickly to the base function (via symbol map in binary file), hence faster.

PR that to the repo and ping Shaneee.

 

[backinblackx86]

smashing

 

[CaseySJ]

PR that to the repo and ping Shaneee.

No worries, I’ll take care of it.

 

[ExtremeXT]

Has anyone been able to inject this in Ventura? I'm still trying, but no success so far.

Yes, it should work by just injecting it with OpenCore.

 

[CaseySJ]

Yes, it should work by just injecting it with OpenCore.

The I225 device got disabled when I tried this yesterday. Will give it another go shortly.

 

[ExtremeXT]

The Gigabyte AM5 user whom I posted the screenshot a few pages back has an ALC897 codec instead of ALC4080 (USB audio) and their audio is 99% perfect with Speedkeeper, I wonder if it's related to the USB audio.

 

[CaseySJ]

The I225 device got disabled when I tried this yesterday. Will give it another go shortly.

Unfortunately no success with AppleIntelI210Ethernet in Ventura. I've tried:

• Raising Probe Score to 4000 (DriverKit-AppleEthernetE2000 is 3000)
• Using e1000=0
• Using device-id = 0xF3158680

 

[CaseySJ]

The Gigabyte AM5 user whom I posted the screenshot a few pages back has an ALC897 codec instead of ALC4080 (USB audio) and their audio is 99% perfect with Speedkeeper, I wonder if it's related to the USB audio.

It's 99% perfect with SpeedKeeper on this system as well. That's what I mean when I say audio from all sources is virtually perfect when SpeedKeeper is enabled.

 

[ExtremeXT]

It's 99% perfect with SpeedKeeper on this system as well. That's what I mean when I say audio from all sources is virtually perfect when SpeedKeeper is enabled.

Could you record a sample coming from the internal audio with and without Speedkeeper?

 

[CaseySJ]

About audio, I found this in an old custom kernel for High Sierra (Screenshot 1). My idea would be to either:
1. Make a kext that calculates the TSC and FSB Frequency values
2. Find an easy way to get these values from the OS and make a script that calculates them. With the proper values, the user can manually input these values into the InitialTSC and FSBFrequency fields in the config.plist
I am positive that it would work if done properly, since it seems to be the same problem as on AMD APUs.
...

With PCI issues resolved, now seems like a good time to reopen the audio discussion.

Let's consider various strategies for tackling this problem before we embark on any particular strategy.

If we agree that obtaining correct FSB and TSC frequencies is the key, then the first few questions are:

• Should we discount or ignore the frequencies reported by sysctl -a | grep tsc and sysctl -a | grep freq ?
• Should we discount or ignore the frequencies reported in OpenCore boot log?
• If we cannot trust the values from both sysctl and OpenCore, how can we still get reliable or usable values from the OS?
• What parameters would we need to obtain from the OS that will allow us to calculate TSC and FSB frequencies?